“To extract the maximum impact and effectiveness of AI requires focus and investment in data; otherwise, the outcome may not be what you thought it would be.”
As we enter the second half of 2025 it is probably right to say we are now living in a world dominated by AI. That’s not to say it has taken over everything we do each day – our taxis are still driven by real people, our kids still get taught in a classroom with a real teacher, we still dress ourselves each morning.
But every aspect of what we do and how we do it can now be seen to be being reviewed – can AI help to make things easier/quicker/simpler/cheaper/more efficient/more effective.
And there is plenty of evidence this approach is accelerating – AI start-ups seem to be the only gig in town when it comes to serious investment in AI-native startups in Europe – in the first half of this year, they hit €3.04bn — 61% more than at the same time in 2024. VCs are spreading their wings to envisage driverless taxis, agent-based education tools and maybe even a Fred Flintstone-style self-dressing contraption (who knows?)
So what does that mean for our industry? AI is clearly becoming more important, even front and centre, for more businesses each day. Our recent research conducted in partnership with EP Hospitality found that 91% of businesses are currently using AI is some guise
What is striking is how this correlates to the RAND research last year, which showed that most internal AI initiatives fail to deliver value, suggesting that up to 80% of projects fail. For the record TechRepublic puts the figure at 85%. Why is this?
The answer must include “data”.
The answer must include “data”. Or to be more accurate, the problem with data. Without perfect and accurate data, no AI model can deliver effective results. LLMs are hooked on data that is available to them, but to be truly transformational and effective, AI programmes typically need data that is not easily found. Which is perhaps why 100% of respondents to the recent survey said more education and insight are required on the way AI can unlock the cost savings promised.
Take pricing – no supplier publishes the products’ listings and prices for each of its customers; that would be madness. Data like this is locked away, hidden on servers, local drives and impregnable (or so they say) ERP systems. Therefore, any application that promises potential efficiency savings, cost reduction, and process improvement needs that data. Without it, effective and game-changing results simply cannot be delivered. What is likely is that an AI programme that originated without thinking of data first will get you to a position where you fail, but you get there faster and with a bigger bang. Not the desired outcome.
So, before you seriously consider any AI-based initiative which requires data, don’t be one of the 80%. As you scope and plan your programme, simply remember that one of the most popular dishes to originate from SE Asia will help you – Perfect Accurate Data, THen AI.
